Matplotlib中disk-cyclide坐标参数化曲面绘图报错求助

尝试绘制采用disk-cyclide坐标参数化的曲面，参数化函数及Lambda函数如图所示。运行代码时出现错误提示“Z必须是二维数组”，但该曲面由三个变量参数化，请问如何调整以满足绘图要求？

size = 250    
psi = np.linspace(0, 2*np.pi, size)
mu = np.linspace(0,K, size)
nu =  np.linspace(0, Kl, size)
Mu, Nu, Psi = np.meshgrid(mu, nu, psi)

k = 0.963
kl = 0.270

K = sp.special.ellipk(k)

Kl = sp.special.ellipk(kl)


Lambda = 1 - (sp.special.ellipj(Mu,k)[2]**2)*(sp.special.ellipj(Nu,kl)[0])

x = (1/Lambda)*(sp.special.ellipj(Mu,k)[1])*(sp.special.ellipj(Nu,kl)[1])*np.cos(Psi)

y = (a/Lambda)*(sp.special.ellipj(Mu,k)[1])*(sp.special.ellipj(Nu,kl)[1])*np.sin(Psi)

z = (a/Lambda)*(sp.special.ellipj(Mu,k)[0])*(sp.special.ellipj(Mu,k)[2])*(sp.special.ellipj(Nu,kl)[0])*(sp.special.ellipj(Nu,kl)[2])

fig = plt.figure()
ax = fig.gca(projection = '3d')

ax.plot_surface(x, y, z, rstride=1, cstride=1, cmap="hot")
plt.show()   

问题根源 #

plot_surface仅接受二维数组作为输入，但你通过meshgrid(mu, nu, psi)生成了三维数组（三个参数维度），这是核心矛盾。disk-cyclide本质是二维曲面，属于2维流形，只需要两个独立参数，三个参数中必然存在依赖关系，或是你误解了参数化逻辑。

调整步骤 #

1. 修正参数维度：保留两个独立参数，比如固定第三个参数为某个值，或选择其中两个参数生成二维网格。
   示例：将mu和psi作为独立参数，固定nu为中间值：

   # 生成二维网格：mu和psi
   size = 250    
   psi = np.linspace(0, 2*np.pi, size)
   mu = np.linspace(0, K, size)
   Mu, Psi = np.meshgrid(mu, psi)
   # 固定nu为Kl/2，也可替换为其他值或遍历多个值生成曲面族
   Nu = np.full_like(Mu, Kl/2)
   

2. 修复未定义变量：代码中y和z用到的a未定义，需补充赋值（比如a = 1.0）。

3. 确保输出为二维数组：重新计算x、y、z，确认最终结果是二维数组后，再传入plot_surface。

完整修正示例代码 #

import numpy as np
import scipy.special as sp
import matplotlib.pyplot as plt

# 补充未定义的参数a
a = 1.0
k = 0.963
kl = 0.270

# 先计算椭圆积分，避免未定义问题
K = sp.ellipk(k)
Kl = sp.ellipk(kl)

size = 250    
# 选择两个独立参数：mu和psi
psi = np.linspace(0, 2*np.pi, size)
mu = np.linspace(0, K, size)
Mu, Psi = np.meshgrid(mu, psi)
# 固定nu为中间值
Nu = np.full_like(Mu, Kl/2)

# 计算Lambda与坐标
Lambda = 1 - (sp.ellipj(Mu, k)[2]**2) * (sp.ellipj(Nu, kl)[0])

x = (1/Lambda) * sp.ellipj(Mu, k)[1] * sp.ellipj(Nu, kl)[1] * np.cos(Psi)
y = (a/Lambda) * sp.ellipj(Mu, k)[1] * sp.ellipj(Nu, kl)[1] * np.sin(Psi)
z = (a/Lambda) * sp.ellipj(Mu, k)[0] * sp.ellipj(Mu, k)[2] * sp.ellipj(Nu, kl)[0] * sp.ellipj(Nu, kl)[2]

fig = plt.figure()
ax = fig.gca(projection='3d')
ax.plot_surface(x, y, z, rstride=1, cstride=1, cmap="hot")
plt.show()

曲面族绘制说明 #

如果disk-cyclide的参数化确实需要三个参数，说明它是带参数的曲面族，此时需要遍历第三个参数，绘制多个曲面：

# 遍历nu的多个值，绘制曲面族
nu_values = np.linspace(0, Kl, 5)
for nu_val in nu_values:
    Nu = np.full_like(Mu, nu_val)
    # 重复计算坐标并绘制
    Lambda = 1 - (sp.ellipj(Mu, k)[2]**2) * (sp.ellipj(Nu, kl)[0])
    x = (1/Lambda) * sp.ellipj(Mu, k)[1] * sp.ellipj(Nu, kl)[1] * np.cos(Psi)
    y = (a/Lambda) * sp.ellipj(Mu, k)[1] * sp.ellipj(Nu, kl)[1] * np.sin(Psi)
    z = (a/Lambda) * sp.ellipj(Mu, k)[0] * sp.ellipj(Mu, k)[2] * sp.ellipj(Nu, kl)[0] * sp.ellipj(Nu, kl)[2]
    ax.plot_surface(x, y, z, rstride=1, cstride=1, cmap="hot", alpha=0.6)

内容的提问来源于stack exchange，提问作者João Renato